A developer built a USB-connected Claude figurine that lights up when Claude Code finishes a response - a hardware fix for one of the more mundane friction points in AI-assisted coding.
The setup is straightforward: a small Claude-branded figurine connects via USB, and a script watches Claude Code's output. When a response completes, the device activates - light, sound, or both - pulling the developer's attention away from whatever else they're doing without requiring them to keep the terminal visible.
The problem it solves is real. Claude Code generates responses that can take anywhere from a few seconds to several minutes depending on task complexity. Most developers deal with this by either watching the terminal the whole time (which means staying at the desk) or periodically tabbing back to check (which wastes time and still misses completions). A physical signal breaks that tradeoff in a way that software hasn't - Claude Code has no native push notification, no completion sound, and no system tray alert.
The figurine uses standard maker components: a USB microcontroller, LED output, and a small monitoring script. The project drew attention because it's reproducible with cheap hardware, and the core pattern (watch for a specific terminal output signal, trigger a USB device) applies to other long-running processes as well - test suites, build pipelines, deployments.
Several developers mentioned building similar setups for CI/CD pipelines and database migrations. The Claude-specific version captures something about the current state of AI-assisted development: these tools are fast enough to be genuinely useful, but some tasks run long enough that context-switching is a real problem - and the tools themselves haven't solved the notification gap.